[PATCH] block: Increase BLK_DEF_MAX_SECTORS_CAP

[PATCH] block: Increase BLK_DEF_MAX_SECTORS_CAP

[Damien Le Moal]

Back in 2015, commit d2be537c3ba3 ("block: bump BLK_DEF_MAX_SECTORS to
2560") increased the default maximum size of a block device I/O to 2560
sectors (1280 KiB) to "accommodate a 10-data-disk stripe write with
chunk size 128k". This choice is rather arbitrary and since then,
improvements to the block layer have software RAID drivers correctly
advertize their stripe width through chunk_sectors and abuses of
BLK_DEF_MAX_SECTORS_CAP by drivers (to set the HW limit rather than the
default user controlled maximum I/O size) have been fixed.

Since many block devices can benefit from a larger value of
BLK_DEF_MAX_SECTORS_CAP, and in particular HDDs, increase this value to
be 4MiB, or 8192 sectors.

Suggested-by: Martin K . Petersen <martin.petersen@oracle.com>
Signed-off-by: Damien Le Moal <dlemoal@kernel.org>
---
 include/linux/blkdev.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h
index 85aab8bc96e7..7c35b2462048 100644
--- a/include/linux/blkdev.h
+++ b/include/linux/blkdev.h
@@ -1238,7 +1238,7 @@ enum blk_default_limits {
  * Not to be confused with the max_hw_sector limit that is entirely
  * controlled by the driver, usually based on hardware limits.
  */
-#define BLK_DEF_MAX_SECTORS_CAP	2560u
+#define BLK_DEF_MAX_SECTORS_CAP	8192u
 
 static inline struct queue_limits *bdev_limits(struct block_device *bdev)
 {
-- 
2.49.0

Re: [PATCH] block: Increase BLK_DEF_MAX_SECTORS_CAP

[Ming Lei]

On Tue, Jun 17, 2025 at 03:34:30PM +0900, Damien Le Moal wrote:
> Back in 2015, commit d2be537c3ba3 ("block: bump BLK_DEF_MAX_SECTORS to
> 2560") increased the default maximum size of a block device I/O to 2560
> sectors (1280 KiB) to "accommodate a 10-data-disk stripe write with
> chunk size 128k". This choice is rather arbitrary and since then,
> improvements to the block layer have software RAID drivers correctly
> advertize their stripe width through chunk_sectors and abuses of
> BLK_DEF_MAX_SECTORS_CAP by drivers (to set the HW limit rather than the
> default user controlled maximum I/O size) have been fixed.
> 
> Since many block devices can benefit from a larger value of
> BLK_DEF_MAX_SECTORS_CAP, and in particular HDDs, increase this value to
> be 4MiB, or 8192 sectors.
> 
> Suggested-by: Martin K . Petersen <martin.petersen@oracle.com>
> Signed-off-by: Damien Le Moal <dlemoal@kernel.org>
> ---
>  include/linux/blkdev.h | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h
> index 85aab8bc96e7..7c35b2462048 100644
> --- a/include/linux/blkdev.h
> +++ b/include/linux/blkdev.h
> @@ -1238,7 +1238,7 @@ enum blk_default_limits {
>   * Not to be confused with the max_hw_sector limit that is entirely
>   * controlled by the driver, usually based on hardware limits.
>   */
> -#define BLK_DEF_MAX_SECTORS_CAP	2560u
> +#define BLK_DEF_MAX_SECTORS_CAP	8192u

The change itself looks good, but the definition should belong to block
layer internal, so why not move it into internal header?


thanks,
Ming

Re: [PATCH] block: Increase BLK_DEF_MAX_SECTORS_CAP

[Martin K. Petersen]

Damien,

> Since many block devices can benefit from a larger value of
> BLK_DEF_MAX_SECTORS_CAP, and in particular HDDs, increase this value
> to be 4MiB, or 8192 sectors.

Reviewed-by: Martin K. Petersen <martin.petersen@oracle.com>

-- 
Martin K. Petersen

Re: [PATCH] block: Increase BLK_DEF_MAX_SECTORS_CAP

[Damien Le Moal]

On 6/17/25 23:52, Ming Lei wrote:
> On Tue, Jun 17, 2025 at 03:34:30PM +0900, Damien Le Moal wrote:
>> Back in 2015, commit d2be537c3ba3 ("block: bump BLK_DEF_MAX_SECTORS to
>> 2560") increased the default maximum size of a block device I/O to 2560
>> sectors (1280 KiB) to "accommodate a 10-data-disk stripe write with
>> chunk size 128k". This choice is rather arbitrary and since then,
>> improvements to the block layer have software RAID drivers correctly
>> advertize their stripe width through chunk_sectors and abuses of
>> BLK_DEF_MAX_SECTORS_CAP by drivers (to set the HW limit rather than the
>> default user controlled maximum I/O size) have been fixed.
>>
>> Since many block devices can benefit from a larger value of
>> BLK_DEF_MAX_SECTORS_CAP, and in particular HDDs, increase this value to
>> be 4MiB, or 8192 sectors.
>>
>> Suggested-by: Martin K . Petersen <martin.petersen@oracle.com>
>> Signed-off-by: Damien Le Moal <dlemoal@kernel.org>
>> ---
>>  include/linux/blkdev.h | 2 +-
>>  1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h
>> index 85aab8bc96e7..7c35b2462048 100644
>> --- a/include/linux/blkdev.h
>> +++ b/include/linux/blkdev.h
>> @@ -1238,7 +1238,7 @@ enum blk_default_limits {
>>   * Not to be confused with the max_hw_sector limit that is entirely
>>   * controlled by the driver, usually based on hardware limits.
>>   */
>> -#define BLK_DEF_MAX_SECTORS_CAP	2560u
>> +#define BLK_DEF_MAX_SECTORS_CAP	8192u
> 
> The change itself looks good, but the definition should belong to block
> layer internal, so why not move it into internal header?

Good point. Now that no driver uses this macro directly, we can do that. I will
send a V2.

-- 
Damien Le Moal
Western Digital Research

Re: [PATCH] block: Increase BLK_DEF_MAX_SECTORS_CAP

[Christoph Hellwig]

On Tue, Jun 17, 2025 at 03:34:30PM +0900, Damien Le Moal wrote:
> Since many block devices can benefit from a larger value of
> BLK_DEF_MAX_SECTORS_CAP, and in particular HDDs, increase this value to
> be 4MiB, or 8192 sectors.
> 
> Suggested-by: Martin K . Petersen <martin.petersen@oracle.com>
> Signed-off-by: Damien Le Moal <dlemoal@kernel.org>
> ---
>  include/linux/blkdev.h | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h
> index 85aab8bc96e7..7c35b2462048 100644
> --- a/include/linux/blkdev.h
> +++ b/include/linux/blkdev.h
> @@ -1238,7 +1238,7 @@ enum blk_default_limits {
>   * Not to be confused with the max_hw_sector limit that is entirely
>   * controlled by the driver, usually based on hardware limits.
>   */
> -#define BLK_DEF_MAX_SECTORS_CAP	2560u
> +#define BLK_DEF_MAX_SECTORS_CAP	8192u

While we're at nitpicking, maybe define this as

	(SZ_4M >> SECTOR_SHIFT)

to make it a bit more readable?

Otherwise this looks good, the odd number was always rather weird.

Re: [PATCH] block: Increase BLK_DEF_MAX_SECTORS_CAP

[Damien Le Moal]

On 6/18/25 13:59, Christoph Hellwig wrote:
> On Tue, Jun 17, 2025 at 03:34:30PM +0900, Damien Le Moal wrote:
>> Since many block devices can benefit from a larger value of
>> BLK_DEF_MAX_SECTORS_CAP, and in particular HDDs, increase this value to
>> be 4MiB, or 8192 sectors.
>>
>> Suggested-by: Martin K . Petersen <martin.petersen@oracle.com>
>> Signed-off-by: Damien Le Moal <dlemoal@kernel.org>
>> ---
>>  include/linux/blkdev.h | 2 +-
>>  1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h
>> index 85aab8bc96e7..7c35b2462048 100644
>> --- a/include/linux/blkdev.h
>> +++ b/include/linux/blkdev.h
>> @@ -1238,7 +1238,7 @@ enum blk_default_limits {
>>   * Not to be confused with the max_hw_sector limit that is entirely
>>   * controlled by the driver, usually based on hardware limits.
>>   */
>> -#define BLK_DEF_MAX_SECTORS_CAP	2560u
>> +#define BLK_DEF_MAX_SECTORS_CAP	8192u
> 
> While we're at nitpicking, maybe define this as
> 
> 	(SZ_4M >> SECTOR_SHIFT)
> 
> to make it a bit more readable?

Sure thing.

> 
> Otherwise this looks good, the odd number was always rather weird.
> 
> 


-- 
Damien Le Moal
Western Digital Research